home / mcp / mcp web tools server

MCP Web Tools Server

Provides web-related MCP tools, including web_scrape, to fetch and render web content via MCP clients.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "surya-madhav-mcp": {
      "command": "python",
      "args": [
        "server.py"
      ]
    }
  }
}

You can harness the MCP Web Tools Server to enable LLMs to interact with web content through a dedicated set of tools. It exposes web-related capabilities that you can connect to from an MCP client, test locally, and integrate with clients like Claude for Desktop for a streamlined web-access workflow.

How to use

Connect your MCP client to the MCP Web Tools Server to access its tools. The primary tool available is web_scrape, which converts a URL to a version prefixed with r.jina.ai and returns the content as markdown. Use this to fetch and present web content within your conversations, summaries, or analyses. Start the server locally, then point your MCP client to the running server to load and execute the tools. Once connected, you can select the web_scrape tool, supply a URL, and receive the rendered markdown content directly in your workflow.

How to install

Prerequisites you need installed before starting: Python and a package manager capable of installing Python dependencies.

Step 1: Clone the server repository and navigate into it.

Step 2: Install dependencies.

pip install -r requirements.txt

Step 3: Run the server using the default stdio transport or with SSE transport for remote listening.

# Run the server with stdio transport (default)
python server.py

# Run the server with SSE transport
python server.py --transport sse --host localhost --port 5000

Step 4: If you prefer using a convenience script, you can start the server or the UI via the provided run script.

# Run the server (stdio) using the script
./run.sh server

# Run the server with SSE transport using the script
./run.sh server --transport sse --host localhost --port 5000

# Run the Streamlit UI using the script
./run.sh ui

Optional manual alternative if you want to start the server directly from Python without the script.

python server.py

# Or start with SSE explicitly
python server.py --transport sse --host localhost --port 5000

Available tools

web_scrape

Converts a URL to use r.jina.ai as a prefix and returns the content in markdown format.